Neural Network Hydrological Modeling for Kemaman Catchment

This paper reports on the evaluation of feed forward back-propagation (FFBP) network, radial basis function network (RBFN), and generalized regression neural network (GRNN) for hydrological modeling of Kemaman watershed in Terengganu. Thirteen (13) meteorological parameters are considered in the input, which includes rainfall, temperature, mean relative humidity and evaporation. The outputs are water levels at four river gauging station. The models were developed and the training results compared in terms of the correlation coefficient and normalized root mean square error. It is shown that the RBFN model is superior over the FFBP and GRNN models, and the performance is sensitive to the various input parameters considered.